Catalytic membrane reactors using solid state membranes for the oxidation or decomposition of various chemical species have been studied and used previously. One potentially valuable use of such reactors is in the production of synthesis gas. See, for example, Cable et al. EP patent application 90305684.4 (published Nov. 28, 1990) and Mazanec et al. U.S. Pat. No. 5,306,411.
Synthesis gas, a mixture of CO and H.sub.2, is widely used as a feedstock in the chemical industry for production of bulk chemicals such as methanol, liquid fuel oxygenates and gasoline. Synthesis gas is currently produced from natural gas, i.e. methane, or other light hydrocarbons by steam reforming. In this technique, natural gas is mixed with steam and heated to high temperatures, and the heated mixture is passed over a catalyst, such as Ni on Al.sub.2 O.sub.3, to form synthesis gas which is then collected. Steam reforming has two major disadvantages. First, the chemical reaction to produce CO and H.sub.2 from steam (H.sub.2 O) and natural gas (CH.sub.4) is endothermic, i.e. the reaction requires energy. Roughly one third of the natural gas consumed in the steam reforming process goes to produce heat to drive the reaction, rather than to produce CO and H.sub.2. Second, the ratio of H.sub.2 :CO in the synthesis gas produced by steam reforming is typically relatively high, from 3:1 up to about 5:1. For most efficient use in the synthesis of methanol, the ratio of H.sub.2 :CO in synthesis gas should be adjusted to 2:1. Adjusting this ratio adds to the cost and complexity of the processing.
In contrast, the use of a catalytic reactor membrane for production of synthesis gas by partial oxidation of natural gas to CO and H.sub.2 overcomes the disadvantages of steam reforming. First, the reaction to produce synthesis gas mediated by the catalytic membrane reactor (CH.sub.4 +1/2O.sub.2 .fwdarw.CO+2H.sub.2) is exothermic, i.e., the reaction gives off heat. The heat produced can then be beneficially used in a cogeneration facility. Second, the synthesis gas produced using a catalytic membrane reactor should have an H.sub.2 :CO ratio of about 2:1. Additional processing steps are eliminated and all the natural gas consumed can be used to produce synthesis gas.
In a catalytic membrane reactor that facilitates oxidation/reduction reactions, a catalytic membrane separates an oxygen-containing gas from a reactant gas which is to be oxidized. Oxygen (O.sub.2) or other oxygen-containing species (for example, NO.sub.x or SO.sub.x) are reduced at one face of the membrane to oxygen anions that are then transported across the membrane to its other face in contact with the reactant gas. The reactant gas, for example methane, is oxidized, for example CH.sub.4 to CO, by the oxygen anions with generation of electrons at the oxidation surface of the membrane.
Materials for membranes in catalytic membrane reactors must be conductors of oxygen anions, and the materials must be chemically and mechanically stable at the high operating temperatures and under the harsh conditions required for reactor operation. In addition, provision must be made in the reactor for electronic conduction to maintain membrane charge neutrality. Membrane materials of most interest are electron conductors, i.e., they conduct electrons.
Oxygen anion conductivity in a material can result from the presence of oxygen anion defects. Defects are deviations from the ideal composition of a specific compound or deviations of atoms from their ideal positions. Of interest for this invention are defects due to loss of oxygen from a compound leading to empty oxygen sites, i.e. oxygen vacancies, in the crystal lattice. A mechanism of oxygen anion conduction is "jumping" of the oxygen anions from site to site. Oxygen vacancies in a material facilitate this "jumping" and thus, facilitate oxygen anion conduction. Oxygen anion defects can be inherent in the structure of a given material of a given stoichiometry and crystal structure or created in a membrane material through reactions between the membrane material and the gas to which it is exposed under the conditions of operation of the catalytic membrane reactor. In a given system with a given membrane material, both inherent and induced defects can occur.
Materials with inherent oxygen anion vacancies are generally preferred. Loss of oxygen from a membrane material by reaction to create vacancies typically has a large effect on the structure of the material. As oxygen is lost, the size of the crystal lattice increases on a microscopic level. These microscopic changes can lead to macroscopic size changes. Because membrane materials are hard, size increases lead to cracking making the membrane mechanically unstable and unusable.
Electronic conductivity in a reactor is necessary to maintain charge neutrality permitting anion conduction through the membrane. It can be achieved by adding an external circuit to a reactor which allows for current flow. U.S. Pat. Nos. 4,793,904, 4,802,958 and 4,933,054 (all of Mazanec et al.) relate to membrane reactors where electronic conductivity is provided by an external circuit. In these patents, the membrane materials, which are compounds with general stoichiometry AO.sub.2, with fluorite structures, such as yttria-stabilized zirconia, exhibit oxygen-anion conductivity.
Electronic conductivity can also be achieved by doping oxygen-anion conducting materials with a metal ion, as illustrated by U.S. Pat. Nos. 4,791,079 and 4,827,071 (both of Hasbun), to generate dual (electrons and oxygen anions) conducting materials. The Hasbun membranes are composed of fluorites doped with transition metals, including titania- and ceria-doped yttria-stabilized zirconia. The disadvantage of this approach is that the dopant metal ions can act as traps for migrating oxygen anions, inhibiting the ionic conductivity of the membrane.
The preferred method for obtaining electronic conductivity is to use membrane materials which inherently possess this property. Dual conducting mixtures can be prepared by mixing an oxygen-conducting material with an electronically-conducting material to form a composite, multi-component, non-single phase material. Problems associated with this method include possible deterioration of conductivity due to reactivity between the different components of the mixture and possible mechanical instability, if the components have different thermal expansion properties.
Cable et al., in European patent application No. 90305684.4 and the corresponding U.S. Pat. No. 5,306,411 of Mazanec at al. report multi-component solid membranes for oxidation/reduction reactions including the production of synthesis gas. The specific multi-phase components are mixtures of an oxygen-conducting material and an electronically conductive material. The oxygen-anion conducting material of the mixture is described as a perovskite ABO.sub.3, including those materials where A and B represent a mixture of more than one metal ion, for example La.sub.a Sr.sub.b O.sub.3, La.sub.a Sr.sub.b FeO.sub.3, La.sub.a Ca.sub.b CoO.sub.3, SrCo.sub.a Fe.sub.b O.sub.3, and Gd.sub.a Sr.sub.b CoO.sub.3, where a and b are numbers and a+b=1. The electronically-conducting material of the mixture is one or more of a variety of metals, metal oxides, metal-doped metal oxides and including mixed metal oxides of a perovskite structure, for example, YBa.sub.2 Cu.sub.3 O.sub.x where x is a number from 6-7. Exemplified multi-component materials include palladium or platinum metal combined with yttria-stabilized zirconia; lanthanum, chromium and magnesium oxides combined with yttria-stabilized zirconia; BMgLaCrO.sub.x combined with yttria-stabilized zirconia and impregnated on its anode side with praseodymium, yttrium and zirconium; and praseodymium-doped indium oxide combined with yttria-stabilized zirconia.
In the same European patent application No. 90305684.4 and U.S. Pat. No. 5,306,411, single-phase, single-component membrane materials, described as exhibiting both oxygen-anion and electronic conductivity, are reported. The specific materials described are mixed metal oxides having a perovskite structure. The perovskite structure is based on that of the mineral perovskite, CaTiO.sub.3. Perovskites have the general formula ABO.sub.3, where A and B are metal ions. The ideal perovskite structure has a cubic lattice in which a unit cell contains metal ions at the corners of the cell, another metal ion in its center and oxygen ions at the midpoints of the cube edges. Examples of single-phase materials given are: LaCoO.sub.3, La.sub.0.6 Sr.sub.0.4 CoO.sub.3, La.sub.0.2 Sr.sub.0.8 CoO.sub.3, YCoO.sub.3, YBa.sub.2 Cu.sub.3 O.sub.x, where x is a number from 6 to 7, La.sub.0.2 Ca.sub.0.8 CoO.sub.3, La.sub.2 Sr.sub.0.8 FeO.sub.3, La.sub.0.2 Sr.sub.0.8 Fe.sub.8 Cr.sub.2 O.sub.3, Gd.sub.0.2 Sr.sub.0.8 CoO.sub.3, and La.sub.0.2 Sr.sub.0.8 Fe.sub.0.8 Cr.sub.0.1 Co.sub.0.1 O.sub.3.
U.S. Pat. No. 5,356,728 of Balachandran et al. also reports the use of mixed metal oxide materials having dual electron and oxygen anion conductivity as ceramic cores in cross-flow reactors. The mixed metal oxide is described as having a perovskite or perovskite-like structure with preferred perovskite structures comprising metals having atomic numbers 4 (Be), 12 (Mg), 20 to 31 (Ca, SC, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga), 38 to 41 (Sr, Y, Zr, Nb) and 56-71 (Ba, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu). Example formulas of oxygen-anion conductive ceramics listed are La.sub.a Sr.sub.b CoO.sub.3, La.sub.a Ca.sub.b CoO.sub.3, La.sub.a Sr.sub.b FeO.sub.3, SrCo.sub.a Fe.sub.b O.sub.3, and Gd.sub.a Sr.sub.b CoO.sub.3, where the sum of a+b is from about 1 to about 1.5. Materials said to be preferred are SrCo.sub.0.5 FeO.sub.x, SrCo.sub.0.8 Fe.sub.0.2 O.sub.x and La.sub.0.2 Sr.sub.0.8 Co.sub.0.4 Fe.sub.0.6 O.sub.x.
WO 94/24065 of Balachandran et al., which takes priority from U.S. Pat. No. 5,356,728, reports crystalline mixed metal oxide compositions of formula Sr.sub..alpha. (Fe.sub.1-x Co.sub.x).sub..alpha.+.beta. where O.sub..delta., where x is a number from 0.1 up to 1, .alpha. is a number from 1 to about 4, .beta. is a number in a range upward from 0 to about 20 and .delta. is a number which renders the compound charge neutral which are useful as membrane materials with oxygen anion conductivity. More specifically the formula for membrane materials is given as Sr.sub.4 (Fe.sub.1-x Co.sub.x).sub.6 O.sub.67 and the composition SrCo.sub.0.5 FeO.sub..delta. is specifically exemplified. The composition is also said to have a characteristic powder X-ray diffraction pattern comprised of principal lines given in Table 1 of the reference.
WO 94/24065 also reports the fabrication of ceramic cores for cross-flow reactors from SrCo.sub.0.8 Fe.sub.0.2 O.sub.x and La.sub.0.2 Sr.sub.0.8 Co.sub.0.4 Fe.sub.0.6 O.sub.x (materials described as preferred in U.S. Pat. No. 5,356,728) and the use of these cores in catalytic reactors for production of synthesis gas. The core made from SrCo.sub.0.8 Fe.sub.0.2 O.sub.x was reported to transport oxygen (0.5 to 3.5 cm.sup.3 /min-cm.sup.2 oxygen permeation rate), but to have fractured after a relatively short time under test conditions. The core made from La.sub.0.2 Sr.sub.0.8 Co.sub.0.4 Fe.sub.0.6 O.sub.x was reported to have fractured in testing without exhibiting oxygen transport.
Teraoka Y., Zhang, H-M., Okamota, K., Yamazoe, N. (1988) Mat. Res. Bull. 23:51-58 and Teraoka, Y., Zhang, H-M., Furukawa, S., Yamazoe, N. (1985) Chemistry Letts. pp. 1743-1746 relate to oxygen permeation and mixed ionic and electronic properties of perovskite-type oxides La.sub.1-x Sr.sub.x Co.sub.1-y Fe.sub.y O.sub.3-.delta.. Teraoka, Y., Nobunaga, T., Yamazoe, N. (1988) Chemistry Lett. pp. 503-506 relates to the effect of cation substitution on the oxygen semipermeability of perovskite-type oxides. Matsumoto, Y., Yamada, S., Nishida, T., Sato E. (1980) J. Electrochem. Soc. 127(11):2360-2364 relates to use of La.sub.1-x Sr.sub.x Fe.sub.1-y Co.sub.y O.sub.3 as electrodes for oxygen evolution reactions in alkaline solution. Goodenough, J. B., Ruiz-Diaz, J. E., Zhen, Y. S. (1990) Solid State Ionics 44:21-31 and Zhen, Y. S., Goodenough J. B. (1990) Mat. Res. Bull. 25:785-790 relate to oxide-ion conduction in Ba.sub.2 In.sub.2 O.sub.5 and Ba.sub.3 In.sub.2 MO.sub.8. Ba.sub.2 In.sub.2 O.sub.5 was shown to have a brownmillerite structure with ordered oxygen vacancies below a given transition temperature where ion conductivity was low. Ordered oxygen vacancies are said to inhibit oxide-ion conductivity.
U.S. Pat. No. 5,397,541 of Post et al. relates to an oxygen sensor which is "based on a thin film of a compound oxide supported on a substrate such as quartz." The oxide has a general formula ABO.sub.2.5+x where x is a variable changing from about 0 to about 0.5 between oxygen-depleted and oxygen-rich forms. A SrFeO.sub.2.5+x +O.sub.2 and a Sr.sub.0.9 La.sub.0.1 FeO.sub.2.5+x +O.sub.2 system are specifically disclosed. The thin film is formed by a laser ablation step from a sintered pellet target of oxygen-rich oxide or oxygen-depleted oxide.